{-# LANGUAGE BangPatterns #-} {-# LANGUAGE MagicHash #-} {- | Copyright: (c) 2016-2019 Artyom Kazak (c) 2019-2020 Kowainik SPDX-License-Identifier: MPL-2.0 Maintainer: Kowainik Note: a lot of these functions are available for other types (in their respective packages): * @@ provides 'indexed' and lots of other functions beginning with “i”. * @@ and @@ provide similar functions, but use a different naming convention (e.g. @@ for maps and @@ for sequences). * provides several typeclasses for indexed functions that work on maps, lists, vectors, bytestrings, and so on (in @@), but unfortunately they are pretty slow for lists. -} module Data.List.Index ( -- * Original functions indexed , deleteAt , setAt , modifyAt , updateAt , insertAt -- * Adapted functions from "Data.List" -- $adapted -- ** Maps , imap , imapM , imapM_ , ifor , ifor_ -- ** Folds , ifoldr , ifoldl , ifoldl' , iall , iany , iconcatMap -- ** Sublists , ifilter , ipartition , itakeWhile , idropWhile -- ** Zipping , izipWith , izipWithM , izipWithM_ -- ** Search , ifind , ifindIndex , ifindIndices -- * Less commonly used functions -- ** Zipping , izipWith3 , izipWith4 , izipWith5 , izipWith6 , izipWith7 -- ** Monadic functions , iforM , iforM_ , itraverse , itraverse_ , ireplicateM , ireplicateM_ , ifoldrM , ifoldlM -- ** Folds , ifoldMap , imapAccumR , imapAccumL ) where import Data.Foldable (sequenceA_) import Data.Maybe (listToMaybe) import Data.Semigroup (Semigroup ((<>))) import GHC.Base (Int (..), Int#, build, oneShot, (+#)) {- Left to do: Functions ~~~~~~~~~ alterF or something? iscanl iscanl' iscanl1 iscanr iscanr1 iiterate? backpermute? minIndex/maxIndex? -} {- | 'indexed' pairs each element with its index. >>> indexed "hello" [(0,'h'),(1,'e'),(2,'l'),(3,'l'),(4,'o')] /Subject to fusion./ -} indexed :: [a] -> [(Int, a)] indexed xs = go 0# xs where go i (a:as) = (I# i, a) : go (i +# 1#) as go _ _ = [] {-# NOINLINE [1] indexed #-} indexedFB :: ((Int, a) -> t -> t) -> a -> (Int# -> t) -> Int# -> t indexedFB c = \x cont i -> (I# i, x) `c` cont (i +# 1#) {-# INLINE [0] indexedFB #-} {-# RULES "indexed" [~1] forall xs. indexed xs = build (\c n -> foldr (indexedFB c) (\_ -> n) xs 0#) "indexedList" [1] forall xs. foldr (indexedFB (:)) (\_ -> []) xs 0# = indexed xs #-} {- | 'deleteAt' deletes the element at an index. If the index is negative or exceeds list length, the original list will be returned. -} deleteAt :: Int -> [a] -> [a] deleteAt i ls | i < 0 = ls | otherwise = go i ls where go 0 (_:xs) = xs go n (x:xs) = x : go (n-1) xs go _ [] = [] {-# INLINE deleteAt #-} {- | 'setAt' sets the element at the index. If the index is negative or exceeds list length, the original list will be returned. -} setAt :: Int -> a -> [a] -> [a] setAt i a ls | i < 0 = ls | otherwise = go i ls where go 0 (_:xs) = a : xs go n (x:xs) = x : go (n-1) xs go _ [] = [] {-# INLINE setAt #-} {- | 'modifyAt' applies a function to the element at the index. If the index is negative or exceeds list length, the original list will be returned. -} modifyAt :: Int -> (a -> a) -> [a] -> [a] modifyAt i f ls | i < 0 = ls | otherwise = go i ls where go 0 (x:xs) = f x : xs go n (x:xs) = x : go (n-1) xs go _ [] = [] {-# INLINE modifyAt #-} {- | 'updateAt' applies a function to the element at the index, and then either replaces the element or deletes it (if the function has returned 'Nothing'). If the index is negative or exceeds list length, the original list will be returned. -} updateAt :: Int -> (a -> Maybe a) -> [a] -> [a] updateAt i f ls | i < 0 = ls | otherwise = go i ls where go 0 (x:xs) = case f x of Nothing -> xs Just x' -> x' : xs go n (x:xs) = x : go (n-1) xs go _ [] = [] {-# INLINE updateAt #-} {- | 'insertAt' inserts an element at the given position: @ (insertAt i x xs) !! i == x @ If the index is negative or exceeds list length, the original list will be returned. (If the index is equal to the list length, the insertion can be carried out.) -} insertAt :: Int -> a -> [a] -> [a] insertAt i a ls | i < 0 = ls | otherwise = go i ls where go 0 xs = a : xs go n (x:xs) = x : go (n-1) xs go _ [] = [] {-# INLINE insertAt #-} {- David Feuer says that drop-like functions tend to have problems when implemented with folds: . I haven't been able to observe this, but since Data.List defines drop/dropWhile/etc that don't fuse, let's do it here as well – just in case. The original version (that does fuse) is below. -- The plan is that if it does inline, it'll be fast; and if it doesn't -- inline, the former definition will be used and sharing will be preserved -- (i.e. if i == 0, it won't rebuild the whole list). deleteAtFB :: Int -> (a -> t -> t) -> a -> (Int# -> t) -> Int# -> t deleteAtFB (I# i) c = \x r k -> case k ==# i of 0# -> x `c` r (k +# 1#) _ -> r (k +# 1#) {-# INLINE [0] deleteAtFB #-} {-# RULES "deleteAt" [~1] forall i xs. deleteAt i xs = build (\c n -> foldr (deleteAtFB i c) (\_ -> n) xs 0#) "deleteAtList" [1] forall i xs. foldr (deleteAtFB i (:)) (\_ -> []) xs 0# = deleteAt i xs #-} -} {- $adapted These functions mimic their counterparts in "Data.List" – 'imap', for instance, works like 'map' but gives the index of the element to the modifying function. Note that left folds have the index argument /after/ the accumulator argument – that's the convention adopted by containers and vector (but not lens). -} {- | /Subject to fusion./ -} imap :: (Int -> a -> b) -> [a] -> [b] imap f ls = go 0# ls where go i (x:xs) = f (I# i) x : go (i +# 1#) xs go _ _ = [] {-# NOINLINE [1] imap #-} imapFB :: (b -> t -> t) -> (Int -> a -> b) -> a -> (Int# -> t) -> Int# -> t imapFB c f = \x r k -> f (I# k) x `c` r (k +# 1#) {-# INLINE [0] imapFB #-} {-# RULES "imap" [~1] forall f xs. imap f xs = build (\c n -> foldr (imapFB c f) (\_ -> n) xs 0#) "imapList" [1] forall f xs. foldr (imapFB (:) f) (\_ -> []) xs 0# = imap f xs #-} {- Note: we don't apply the *FB transformation to 'iconcatMap' because it uses 'ifoldr' instead of 'foldr', and 'ifoldr' might get inlined itself, and rewriting 'iconcatMap' with 'foldr' instead of 'ifoldr' is annoying. So, in theory it's a small optimisation possibility (in practice I'm not so sure, given that functions with 'build' don't seem to perform worse than functions without it). -} iconcatMap :: (Int -> a -> [b]) -> [a] -> [b] iconcatMap f xs = build $ \c n -> ifoldr (\i x b -> foldr c b (f i x)) n xs {-# INLINE iconcatMap #-} ifoldMap :: (Semigroup m, Monoid m) => (Int -> a -> m) -> [a] -> m ifoldMap p ls = foldr go (\_ -> mempty) ls 0# where go x r k = p (I# k) x <> r (k +# 1#) {-# INLINE ifoldMap #-} {- | /Subject to fusion./ -} iall :: (Int -> a -> Bool) -> [a] -> Bool iall p ls = foldr go (\_ -> True) ls 0# where go x r k = p (I# k) x && r (k +# 1#) {-# INLINE iall #-} {- | /Subject to fusion./ -} iany :: (Int -> a -> Bool) -> [a] -> Bool iany p ls = foldr go (\_ -> False) ls 0# where go x r k = p (I# k) x || r (k +# 1#) {-# INLINE iany #-} imapM :: Monad m => (Int -> a -> m b) -> [a] -> m [b] imapM f as = ifoldr k (return []) as where k i a r = do x <- f i a xs <- r return (x:xs) {-# INLINE imapM #-} iforM :: Monad m => [a] -> (Int -> a -> m b) -> m [b] iforM = flip imapM {-# INLINE iforM #-} itraverse :: Applicative m => (Int -> a -> m b) -> [a] -> m [b] itraverse f as = ifoldr k (pure []) as where k i a r = (:) <$> f i a <*> r {-# INLINE itraverse #-} ifor :: Applicative m => [a] -> (Int -> a -> m b) -> m [b] ifor = flip itraverse {-# INLINE ifor #-} {- | /Subject to fusion./ -} imapM_ :: Monad m => (Int -> a -> m b) -> [a] -> m () imapM_ f as = ifoldr k (return ()) as where k i a r = f i a >> r {-# INLINE imapM_ #-} {- | /Subject to fusion./ -} iforM_ :: Monad m => [a] -> (Int -> a -> m b) -> m () iforM_ = flip imapM_ {-# INLINE iforM_ #-} {- | /Subject to fusion./ -} itraverse_ :: Applicative m => (Int -> a -> m b) -> [a] -> m () itraverse_ f as = ifoldr k (pure ()) as where k i a r = f i a *> r {-# INLINE itraverse_ #-} {- | /Subject to fusion./ -} ifor_ :: Applicative m => [a] -> (Int -> a -> m b) -> m () ifor_ = flip itraverse_ {-# INLINE ifor_ #-} {- | Perform a given action @n@ times. Behaves like @for_ [0..n-1]@, but avoids . If you want more complicated loops (e.g. counting downwards), consider the package. -} ireplicateM :: Applicative m => Int -> (Int -> m a) -> m [a] ireplicateM cnt f = go 0 where go !i | i >= cnt = pure [] | otherwise = (:) <$> f i <*> go (i + 1) {-# INLINE ireplicateM #-} {- | NB. This function intentionally uses 'Monad' even though 'Applicative' is enough. That's because the @transformers@ package didn't have an optimized definition of ('*>') for 'StateT' prior to 0.5.3.0, so for a common case of 'StateT' this function would be 40 times slower with the 'Applicative' constraint. -} ireplicateM_ :: Monad m => Int -> (Int -> m a) -> m () ireplicateM_ cnt f = if cnt > 0 then go 0 else return () where -- this is 30% faster for Maybe than the simpler -- go i | i == cnt = return () -- | otherwise = f i >> go (i + 1) cnt_ = cnt-1 go !i = if i == cnt_ then f i >> return () else f i >> go (i + 1) {-# INLINE ireplicateM_ #-} -- Using unboxed ints here doesn't seem to result in any benefit ifoldr :: (Int -> a -> b -> b) -> b -> [a] -> b ifoldr f z xs = foldr (\x g i -> f i x (g (i+1))) (const z) xs 0 {-# INLINE ifoldr #-} ifoldrM :: Monad m => (Int -> a -> b -> m b) -> b -> [a] -> m b ifoldrM f z xs = ifoldr k (return z) xs where k i a r = f i a =<< r {-# INLINE ifoldrM #-} imapAccumR :: (acc -> Int -> x -> (acc, y)) -> acc -> [x] -> (acc, [y]) imapAccumR f z xs = foldr (\x g i -> let (a, ys) = g (i+1) (a', y) = f a i x in (a', y:ys)) (const (z, [])) xs 0 {-# INLINE imapAccumR #-} {- ifoldr1 :: (Int -> a -> a -> a) -> [a] -> a ifoldr1 f = go 0# where go _ [x] = x go i (x:xs) = f (I# i) x (go (i +# 1#) xs) go _ [] = errorEmptyList "ifoldr1" {-# INLINE [0] ifoldr1 #-} -} {- | The index isn't the first argument of the function because that's the convention adopted by containers and vector (but not lens). /Subject to fusion./ -} ifoldl :: forall a b. (b -> Int -> a -> b) -> b -> [a] -> b ifoldl k z0 xs = foldr (\(v::a) (fn :: (Int, b) -> b) -> oneShot (\((!i)::Int, z::b) -> fn (i+1, k z i v))) (snd :: (Int, b) -> b) xs (0, z0) {-# INLINE ifoldl #-} {- | /Subject to fusion./ -} ifoldl' :: forall a b. (b -> Int -> a -> b) -> b -> [a] -> b ifoldl' k z0 xs = foldr (\(v::a) (fn :: (Int, b) -> b) -> oneShot (\((!i)::Int, z::b) -> z `seq` fn (i+1, k z i v))) (snd :: (Int, b) -> b) xs (0, z0) {-# INLINE ifoldl' #-} {- | /Subject to fusion./ -} ifoldlM :: Monad m => (b -> Int -> a -> m b) -> b -> [a] -> m b ifoldlM f z xs = ifoldl k (return z) xs where k a i r = do a' <- a; f a' i r {-# INLINE ifoldlM #-} imapAccumL :: (acc -> Int -> x -> (acc, y)) -> acc -> [x] -> (acc, [y]) imapAccumL f z xs = foldr (\(x::a) (r :: (Int,acc) -> (acc,[y])) -> oneShot (\((!i)::Int, s::acc) -> let (s', y) = f s i x (s'', ys) = r (i+1, s') in (s'', y:ys))) ((\(_, a) -> (a, [])) :: (Int,acc) -> (acc,[y])) xs (0, z) {-# INLINE imapAccumL #-} {- ifoldl1 :: (a -> Int -> a -> a) -> [a] -> a ifoldl1 f (x:xs) = ifoldl f x xs ifoldl1 _ [] = errorEmptyList "ifoldl1" ifoldl1' :: (a -> Int -> a -> a) -> [a] -> a ifoldl1' f (x:xs) = ifoldl' f x xs ifoldl1' _ [] = errorEmptyList "ifoldl1'" -} ifilter :: (Int -> a -> Bool) -> [a] -> [a] ifilter p ls = go 0# ls where go i (x:xs) | p (I# i) x = x : go (i +# 1#) xs | otherwise = go (i +# 1#) xs go _ _ = [] {-# NOINLINE [1] ifilter #-} ifilterFB :: (a -> t -> t) -> (Int -> a -> Bool) -> a -> (Int# -> t) -> Int# -> t ifilterFB c p = \x r k -> if p (I# k) x then x `c` r (k +# 1#) else r (k +# 1#) {-# INLINE [0] ifilterFB #-} {-# RULES "ifilter" [~1] forall p xs. ifilter p xs = build (\c n -> foldr (ifilterFB c p) (\_ -> n) xs 0#) "ifilterList" [1] forall p xs. foldr (ifilterFB (:) p) (\_ -> []) xs 0# = ifilter p xs #-} itakeWhile :: (Int -> a -> Bool) -> [a] -> [a] itakeWhile p ls = go 0# ls where go i (x:xs) | p (I# i) x = x : go (i +# 1#) xs | otherwise = [] go _ _ = [] {-# NOINLINE [1] itakeWhile #-} itakeWhileFB :: (a -> t -> t) -> (Int -> a -> Bool) -> t -> a -> (Int# -> t) -> Int# -> t itakeWhileFB c p n = \x r k -> if p (I# k) x then x `c` r (k +# 1#) else n {-# INLINE [0] itakeWhileFB #-} {-# RULES "itakeWhile" [~1] forall p xs. itakeWhile p xs = build (\c n -> foldr (itakeWhileFB c p n) (\_ -> n) xs 0#) "itakeWhileList" [1] forall p xs. foldr (itakeWhileFB (:) p []) (\_ -> []) xs 0# = itakeWhile p xs #-} idropWhile :: (Int -> a -> Bool) -> [a] -> [a] idropWhile p ls = go 0# ls where go i (x:xs) | p (I# i) x = go (i +# 1#) xs | otherwise = x:xs go _ [] = [] {-# INLINE idropWhile #-} ipartition :: (Int -> a -> Bool) -> [a] -> ([a],[a]) ipartition p xs = ifoldr (iselect p) ([],[]) xs {-# INLINE ipartition #-} iselect :: (Int -> a -> Bool) -> Int -> a -> ([a], [a]) -> ([a], [a]) iselect p i x ~(ts,fs) | p i x = (x:ts,fs) | otherwise = (ts, x:fs) ifind :: (Int -> a -> Bool) -> [a] -> Maybe (Int, a) ifind p ls = go 0# ls where go i (x:xs) | p (I# i) x = Just (I# i, x) | otherwise = go (i +# 1#) xs go _ _ = Nothing {-# INLINE ifind #-} ifindIndex :: (Int -> a -> Bool) -> [a] -> Maybe Int ifindIndex p = listToMaybe . ifindIndices p ifindIndices :: (Int -> a -> Bool) -> [a] -> [Int] ifindIndices p ls = go 0# ls where go _ [] = [] go i (x:xs) | p (I# i) x = I# i : go (i +# 1#) xs | otherwise = go (i +# 1#) xs {-# NOINLINE [1] ifindIndices #-} ifindIndicesFB :: (Int -> t -> t) -> (Int -> a -> Bool) -> a -> (Int# -> t) -> Int# -> t ifindIndicesFB c p = \x r k -> if p (I# k) x then I# k `c` r (k +# 1#) else r (k +# 1#) {-# INLINE [0] ifindIndicesFB #-} {-# RULES "ifindIndices" [~1] forall p xs. ifindIndices p xs = build (\c n -> foldr (ifindIndicesFB c p) (\_ -> n) xs 0#) "ifindIndicesList" [1] forall p xs. foldr (ifindIndicesFB (:) p) (\_ -> []) xs 0# = ifindIndices p xs #-} {- errorEmptyList :: String -> a errorEmptyList fun = error ("Data.List.Index." ++ fun ++ ": empty list") -} {- | /Subject to fusion in the first argument./ -} izipWith :: (Int -> a -> b -> c) -> [a] -> [b] -> [c] izipWith fun xs ys = go 0# xs ys where go i (a:as) (b:bs) = fun (I# i) a b : go (i +# 1#) as bs go _ _ _ = [] {-# NOINLINE [1] izipWith #-} izipWithFB :: (c -> t -> t) -> (Int -> a -> b -> c) -> a -> b -> (Int# -> t) -> Int# -> t izipWithFB c fun = \x y cont i -> fun (I# i) x y `c` cont (i +# 1#) {-# INLINE [0] izipWithFB #-} {-# RULES "izipWith" [~1] forall f xs ys. izipWith f xs ys = build (\c n -> foldr2 (izipWithFB c f) (\_ -> n) xs ys 0#) "izipWithList" [1] forall f xs ys. foldr2 (izipWithFB (:) f) (\_ -> []) xs ys 0# = izipWith f xs ys #-} -- Copied from GHC.List foldr2 :: (a -> b -> c -> c) -> c -> [a] -> [b] -> c foldr2 k z = go where go [] _ys = z go _xs [] = z go (x:xs) (y:ys) = k x y (go xs ys) {-# INLINE [0] foldr2 #-} foldr2_left :: (a -> b -> c -> d) -> d -> a -> ([b] -> c) -> [b] -> d foldr2_left _k z _x _r [] = z foldr2_left k _z x r (y:ys) = k x y (r ys) {-# RULES "foldr2/left" forall k z ys (g::forall b.(a->b->b)->b->b) . foldr2 k z (build g) ys = g (foldr2_left k z) (\_ -> z) ys #-} izipWith3 :: (Int -> a -> b -> c -> d) -> [a] -> [b] -> [c] -> [d] izipWith3 fun = go 0# where go i (a:as) (b:bs) (c:cs) = fun (I# i) a b c : go (i +# 1#) as bs cs go _ _ _ _ = [] {-# INLINE izipWith3 #-} izipWith4 :: (Int -> a -> b -> c -> d -> e) -> [a] -> [b] -> [c] -> [d] -> [e] izipWith4 fun = go 0# where go i (a:as) (b:bs) (c:cs) (d:ds) = fun (I# i) a b c d : go (i +# 1#) as bs cs ds go _ _ _ _ _ = [] {-# INLINE izipWith4 #-} izipWith5 :: (Int -> a -> b -> c -> d -> e -> f) -> [a] -> [b] -> [c] -> [d] -> [e] -> [f] izipWith5 fun = go 0# where go i (a:as) (b:bs) (c:cs) (d:ds) (e:es) = fun (I# i) a b c d e : go (i +# 1#) as bs cs ds es go _ _ _ _ _ _ = [] {-# INLINE izipWith5 #-} izipWith6 :: (Int -> a -> b -> c -> d -> e -> f -> g) -> [a] -> [b] -> [c] -> [d] -> [e] -> [f] -> [g] izipWith6 fun = go 0# where go i (a:as) (b:bs) (c:cs) (d:ds) (e:es) (f:fs) = fun (I# i) a b c d e f : go (i +# 1#) as bs cs ds es fs go _ _ _ _ _ _ _ = [] {-# INLINE izipWith6 #-} izipWith7 :: (Int -> a -> b -> c -> d -> e -> f -> g -> h) -> [a] -> [b] -> [c] -> [d] -> [e] -> [f] -> [g] -> [h] izipWith7 fun = go 0# where go i (a:as) (b:bs) (c:cs) (d:ds) (e:es) (f:fs) (g:gs) = fun (I# i) a b c d e f g : go (i +# 1#) as bs cs ds es fs gs go _ _ _ _ _ _ _ _ = [] {-# INLINE izipWith7 #-} izipWithM :: Applicative f => (Int -> a -> b -> f c) -> [a] -> [b] -> f [c] izipWithM f as bs = sequenceA (izipWith f as bs) {-# INLINE izipWithM #-} izipWithM_ :: Applicative f => (Int -> a -> b -> f c) -> [a] -> [b] -> f () izipWithM_ f as bs = sequenceA_ (izipWith f as bs) {-# INLINE izipWithM_ #-}